An improved spectrum correlation time-frequency analysis method and its application in fault diagnosis of rolling element bearing

Hongchao Wang,Wenliao Du

doi:10.21595/jve.2020.20888

Abstract

The spectrum correlation (SC) is an effective fault feature extraction method for rolling bearing which is based on second order cyclic statistic. However, the effectiveness of SC will be compromised greatly when the rolling bearing fault feature signal is interfered by noise. To solve the problem, the SC method is improved and the improved spectrum correlation (ISC) method is proposed in the paper, and the noise-resistance virtue of ISC compared with SC is verified through the accelerated fatigue and compound fault test of rolling element bearing. Besides, the vibration signal of fault rolling bearing takes on modulation phenomenon, and extracting the fault characteristic frequency (FCF) or cyclic modulation frequency (CMF) is enough for the purpose of fault diagnosis, and the modulation frequency is neglected usually. However, the extraction result of ISC is not intuitive enough because it extracts the FCF and modulation frequency with its harmonic synchronously. To improve the intuitive feature extraction effect of ISC, the ISC method is improved further in the paper and the integrated improved spectrum correlation (IISC) is proposed which would only extract the FCF or CMF, so much clear and better extraction effectiveness could be obtained by IISC method, and the effectiveness and better fault extraction results by applying IISC method on vibration data of rolling bearing accelerated fatigue and compound fault test are also presented.

Highlights

The failure of rolling element bearing represents a high percentage of breakdowns in rotating machinery [1], and it is meaningful to study effective fault diagnosis method of it in ensuring the reliable running of machinery
The improved spectrum correlation (ISC) method is further improved and the integrated improved spectrum correlation (IISC) method is proposed in the paper which has the advantage of more intuitive feature extraction results compared with ISC, because the IISC is only to extract the fault characteristic frequency (FCF) or cyclic modulation frequency (CMF) and ignore the modulation frequency with its harmonics
It is verified that the Envelope demodulation (ED) method is not effective enough to extract the rolling bearing FCFs when compound fault arises in rolling element bearing

Summary

Introduction

The failure of rolling element bearing represents a high percentage of breakdowns in rotating machinery [1], and it is meaningful to study effective fault diagnosis method of it in ensuring the reliable running of machinery. A new fault diagnosis method of rolling bearing based on principal component analysis and broad learning system was proposed in paper [2]. AN IMPROVED SPECTRUM CORRELATION TIME-FREQUENCY ANALYSIS METHOD AND ITS APPLICATION IN FAULT DIAGNOSIS OF ROLLING ELEMENT BEARING. The FFT spectral analysis result of the signal shown in Eq (8) is presented in Fig. 1 where f represents carrier frequency, and the frequencies of the modulation signal locate symmetrically from f with the interval of cyclic frequency α. In the process of rolling bearing fault diagnosis, it is enough to extract the modulated frequency and ignore the modulated frequency with its harmonic components, so the ISC method is further improved, and IISC method is proposed here: IISC (α, ∆f) = ISC∆ (f, α)df

Rolling bearing accelerated fatigue experiment

Simulation of rolling bearing compound fault

Experiment of rolling bearing compound fault

Findings

Conclusions